DE3233290A1 - DEVICE FOR EXHAUST GAS RECIRCULATION IN PARTICULAR A INTERNAL COMBUSTION ENGINE WORKING WITH ITSELF - Google Patents

Description

March 31, 1932 Mü /? i

ROBERT 3 per CH GMBH, TOGO STUTTGART 1

iinr i: htung for exhaust gas recirculation in an internal combustion engine that operates in particular with compression ignition

State of the art

From US-PS h 177 777 an exhaust gas recirculation control system is known in which the appropriate amount of air from the internal combustion engine is measured by means of an air flow meter and optionally together with further operating savings at et er η eir. Control signal is formed that de α control pressure of a pneumatically actuated exhaust gas recirculation valve changes uni so that controls the amount of recirculated Ac gas. Furthermore, US Pat. No. '777 discloses an exhaust gas recirculation system which si3t the exhaust gas recirculation quantity and compares it with a setpoint value formed from the air quantity signal and changes the setting of the exhaust gas recirculation valve in accordance with the resulting correction signal. In the first example mentioned, there is the laughing part that the exhaust gas recirculation rate is only controlled without any incorrect influences being detected

can. The second solution mentioned requires an arrangement for the amount of exhaust gas actually recirculated. The problems that arise here are well known. They are mainly due to the high temperature ies Exhaust gas and its solids content, so that the exhaust gas measurement extremely in terms of long-term stability is problematic.

Furthermore, the prior art includes systems according to US Patents 4 10U 032, U i £ U 206, h i! +2 1 * 93, h 1T3 203, k 1o1 929, U 250 706, as well as DE- 03 28 51 ISO.

Advantages of the invention

The device according to the invention for exhaust gas recirculation in the case of one that works in particular with compression ignition Internal combustion engine with the features of Main claim is characterized by a comprehensive Signal processing to control the exhaust gas recirculation valve and thus enables an exact and sensitive Quantity control. Furthermore, 3i: h gives a Reduction of switching processes for solenoid valves, which is desirable in view of a low susceptibility to failure of the device.

further gate parts and appropriate configurations of the Invention emerge in connection with the following description of an exemplary embodiment.

drawing

An embodiment of the invention is shown in the drawing and is described in more detail below

J C- V

and explained ;. Figure 1 shows a rough overview a device for iie exhaust gas recirculation in an internal combustion engine with compression ignition, FIG 2 is a block diagram of the control circuit for exhaust gas recirculation, Figure 3 pulse patterns sum explaining the control signals for the two solenoid valve windings as well as the limits of the controller, figure k the signal curve of the valve simulation signal with the boundary areas and Figure 5 is a circuit diagram of the controller with the following three-point switch as well as the output stage control.

Description of the embodiment

The embodiment relates to a device for exhaust gas recirculation in a 3r enakraf t machine operating with compression ignition. This internal combustion engine bears the reference number 10 and an air intake pipe is indicated by 11 and an exhaust pipe by 12. The control of the exhaust gas recirculation is a valve 13 whose Fosition is controllable by means of an adjusting device \\. It comprises a housing 15 with a membrane Is located therein, which is held in a defined position by means of a spring 17. The spring force is superimposed by the effect of the pressure difference between the two chambers of the actuating device, the chamber with the spring 1? is coupled to a pneumatic line 13. Two solenoid valves 19 and 20 enable the pneumatic line 13 to be connected once to the atmosphere and once to a vacuum source, represented for example by a vacuum pump 21. Both solenoid valves 1Q and 2C receive control pulses from a control unit that processes different input variables.

1-

- y-

Sin accelerator pedal position sensor is marked with 22, a Spray duration sensor with 23, a rotary steel sensor with ~ 2 ^, an air volume sensor with 25 and a temperature sensor with 26th Ben. Speed sensor 2m- follows a Jreauer.:- Span.nungs-Wan.dler 28 and the spray duration sensor 23 a Injection duration signal evaluation unit 29 · It comprises a Characteristic curve or a map to at its output Signal with the value

m (SD - SDO)

to be able to leave. 'SL means the measured spray duration, SDO a constant and m a multiplicative Factor.

An acceleration detection stage 30 is connected downstream of the accelerator pedal sensor 22. On the output side, both units or stages 29 and 30 are connected to a summation point 31, which in turn leads to a family of characteristics 32. The so-called lambda characteristic contained therein is determined based on the measured injection duration and thus the measured, learned combustion chamber force substance quantity a basic air quantity signal 'Js, which can be changed in a subsequent difference point 3 ^ depending on the output signal of a speed correction stage 35 and' then represents a signal relating to the air quantity setpoint Usn is compared at a difference point 37 and the relevant difference value ^ χ is then available as an input variable for the PID controller with three-point switch 3δ. As an additional input signal, block 38 receives a signal from temperature sensor 26. On the output side, two lines 33 and 'where unmit ~ el-

3733290 ..- ■,:. ; _v λ /,

bar to the excitation windings 41 and U2 of the two solenoid valves 2 0 and 19.

The basic idea of the arrangement shown in FIG. 1 is it, based on the measured amount of injected fuel an air volume target value "to form, and this target value via the degree of exhaust gas admixture to the To regulate the fresh air proportion.

FIG. 2 shows a block diagram of the control circuit for the exhaust gas recirculation in the subject of FIG. 1. 37 again identifies the comparative parts for the setpoint and actual value. For reasons of expediency with regard to a special embodiment, the reciprocal setpoint and actual value of the air volume are used here. The control deviation ό x determined in this comparison point 37 reaches the unit 38 with a PID controller -5, which has limits, and then to a three-point switch Π6. This is followed by a block • + 5 , which shows the signal behavior of the. Characterized solenoid valves and the membrane interlocking. This block hQ basically identifies a signal position converter which, in the arrangement shown in FIG. 1, has integral behavior in the worst case. In the exhaust gas recirculation valve 13, the valve position; c and the recirculated exhaust gas quantity are assigned. The higher the amount of recirculated exhaust gas, the smaller the proportion of fresh air and vice versa, due to the pressure conditions in the intake pipe. This is shown using the characteristic curve QL over the stroke of the exhaust gas recirculation valve 13. The control loop closes via the air volume sensor 25, which, if configured accordingly, supplies an output signal proportional to the reciprocal value of the air volume and thus supplies an air energy actual signal to the comparative part 37 Provides. That

The corresponding setpoint signal corresponds to the value ^r Jsri shown in FIG. 1 and the 3olL actual value -3 difference forms the input variable of the PID controller * 5 as the Hegel deviation Δ.

The hallmark of the 4Ö three-point switch is its I; ::: re and its hysteresis, the dimensions of which are explained in more detail below. In addition, FIG. 2 shows an actuator simulation unit 50, the input signal of which corresponds to the control signal of the "two irreversible windings 41 and h-2 ίϊγ solenoid valves 19 and 20 and additionally feeds a delayed signal to the input of the three-point switch k6 on the output side," flexible feedback ). -

The PID controller ^ 5 works linearly and only within the dead zone of the following three-point switch h-6. In the event of a small control deviation, this controller also determines the switching frequency of the individual solenoid valves. Within the dead zone or the hysteresis, the two-point switch k6 does not change the value of the signal box. If the output signal of the PID controller * + 5 exceeds a certain value to one side, then one of the two solenoid valves "9, 20" switches on and the diaphragm 16 of the steep device * is shifted with a time constant TI. This results in the behavior of the Blocks kd in Figure 2 as!

Purpose of the hysteresis of the three-point switching ers -o is to avoid a permanent switching back and forth of the valves and thereby reduce their consumption of ier "Jnterdruckversorgung and the service life: u increase the size of the hysteresis is determined by the pre-

■. ■ - ^ -180 4 ο

- r- ^M

ha η ie η. en Z t 5 r u η g s b a η d. The dead zone results from ier desired. Accuracy; and the P amplification of the Controller Up.

The feedback of the three-point roch switch output signal represents an auxiliary rule variable. To improve the Stability, it simulates the reaction of the actuator according to the behavior of an observer. You prevented on this Way a strong overshooting reaction of the control loop. The unit 50 does not generate any PID behavior, rather, in the event that only a minimal correction is necessary, it determines with the release time con-3 ".aa: er. The associated minimum duration of the actuating impulses. The feedback decay time constant is greater, for lower switching frequency.

The time behavior of the PID controller ^ 5 has no effect directly on the actuator, i.e. the valve 13, but it influences the small signal behavior and the Switching frequency. It is therefore particularly important how you react after a switching threshold is exceeded will. If e.g. 3. within the dead zone of the I-controller there is a control deviation that is just still present integrated, then only a smallest possible Affect actuating pulse.

In order to use the controller in connection with a switching point keep close to his optimal working area, is the I-part shortly after crossing a switching, threshold limited. In the case of larger control deviations, the.? And D sections can be limited up to a later date keep walking. This is shown in detail in Figure 3, from which the signal behavior of the PIZ controller -? with its limitations as well as that The behavior of the three-point switch s! + B in detail is shown.

3733290 ν- λ ~ _Λ . ,

-X-

The definition of a minimum and a maximum pulse length is expediently based on the two FIGS . 3 and h. explained.

The valve does not open mechanically when the pulse length is less than timin. The following applies to the minimum effective pulse duration. With a fixed hysteresis, the time constant and the weighting of the feedback (via the actuator simulation unit 50) ro are defined so that the feedback voltage just reaches the value of the hysteresis after a duration of t min. See the corresponding entries in Figure 3, h. In the case of a minimal control deviation, the smallest possible correction has a great effect. 7, ^r hen iie control deviation becomes a larger value, it towers, depending on the size of the? Amplification outside the dead zone, the returned voltage. Ss are then switched accordingly longer pulses.

With an absolute limitation of the controller U5, a maximum pulse length t _T max can be determined. It has to be chosen so small that it is not yet sufficient for the existence of permanent oscillations. The pulse lengths are therefore limited to the range between t min and * _T aa: c and are based on the size of the control deviation.

Sin embodiment of block 33. in Figure *. and is shown in FIG. 5, whereby it should be noted that FIG. 5 again shows two excitation windings U 1 and k2 belonging to solenoid valves, while FIG. 2 shows only one output line with the signal UMV for the sake of simplicity.

^ I au pt bestanat = il 4 Sehalttungsanorinung shown in Figure 5 ies Blocks 30 of Figure 2 are a PID controller - <5 with its limits, a three point he switches with two comparators, a flexible switching pulse feedback to represent the actuator -Simulation unit u.ii ultimately two driver stages for the two control signals of the excitation windings 41 and k2.

The following structure results in detail.

Located between the air volume sensor 25 and the difference point 37 a D element consisting of a series circuit of capacitor 55 and resistor 5o and one of these Series connection parallel · Resistance stood 57. The Air volume setpoint is via a resistor 58 to the Difference point 37 fed in.

The main component of the PID controller is an operational amplifier 00, in whose negative feedback branch there is a capacitor 51. 'Between two bat ¹ ; he voltage lines c2 and 04 are two single-stage voltage dividers with resistors z ~ and 66, 6 "and 63 as well as a multi-stage voltage divider from five resistors 69 to 73. The subtraction point 37 is led directly to the minus input of the operational amplifier 6.1. This minus -Ξ Ingang stand ¹ : additionally connected via a diode 75 to the wiper of the resistor 79, which is designed as a potentiometer, and also via a diode 7o to the wiper of the resistor 73, which is also designed as a potentiometer. In addition, there is a series connection of resistor 77 and capacitor 73 between the negative input and the connection point of the two resistors 73 and 72 as well as a Wi of the 3tan ds capacitor-diode combination between the minus input and the connection points of the resistors c? and 70 as well as 71 and 72. There is a resistor

- yS -

stood 79 one. Capacitor 8θ parallel and this parallel circuit is in turn coupled via a respective diode 31 ■ and with the junctions of the resistors 69 and "3 as well as 71 and 72 Ausgangssei. *; Ig of the operational amplifier 60 is connected to the junction of resistors 70 and 71, which also represents the output of the PID controller -4-5 with its limits. It follows via a resistor 3u the differential point at the input of the three-point switch h6 of Figure 1, at the output of which the signal Λ χ occurs. This point is the negative input of an operational amplifier 35, whose positive input mi * is coupled to the junction of the two resistors 67 and 53. The operational amplifier 35 is counter-coupled by means of a resistor 36.

The three-point switch ± 0 of Figure 2 is formed in the subject of Figure 5 from two threshold switches 33 and 59. A two-stage voltage divider with the three resistors 90 to 92 generates the echo threshold values for the two comparators 88 and 39, with a resistor 93 from the junction of the two resistors 90 and 91 to the negative input of the threshold switch 88 and from the junction of the two resistors 91 and 92 a resistor 9 ^ 2um plus input of the threshold switch 39 leads. The other two inputs of the threshold switches 33 and 39 are each connected to the output of the operational amplifier 35 via a resistor 96 and 9 ″. A capacitor 93 and 99 is also connected between the input terminals of the two threshold switches 88 and 39 Resistances 100 and 101. On the output side, the two threshold value switches 83 and 89 each have a resistor, 32 uni

3733290

ι "j3 with the positive voltage line 6? in connection and also via a resistor 10" and 105 21 with the jate connections of two field effect transistors I06 and 107. While their source connections are each grounded, are between the drain connections and ier plus line o3 dia excitation windings i + 1 and \ 2. of solenoid valves 19 and 20.

A resistor-capacitor-diode network is used to implement the actuator simulation (unit 50 of the subject matter of FIG. 2). A series connection of resistor 110 and diode 111 leads from the output of threshold switch 88 and a series connection of resistor 112 and diode 113 also leads from the drain connection of field effect transistor 107 to a 7-point c indur.gpunkt 114 on which the signal xr from FIG occurs and is routed to the minus input of the operational amplifier 35 via an impedance converter with a 2 'operational amplifier 115 and a resistor 116. In addition, h is located between connecting point 11 and the voltage divider consisting of two resistors 117 and 113, a parallel connection resistor .One II9 and a Kondensaters 120th

The circuit arrangement shown in Figure 5 offers as such, no particular difficulties. Worth mentioning is just the following:

The operational amplifier 60 with its PID circuit receives a limitation of the integral part if its Output voltage to one side- such a difference value has that the diodes 31 or 32 conduct according to the voltage divider 69 and 70 or 71 to 73

3733290

tend to be. Then the branch with the capacitor SO and the resistor 79 switches in parallel to the PI-Z ve ig 77, 73, so that a delayed P controller is created, with a ^{gain corresponding to the resistor T} 9 reduced. The capacitor Su parallel to the resistor 79 dampens the controller against continuous oscillations which could disrupt the exhaust gas recirculation at low speed. The integration time constant (C78) depends on the required control speed.

Bring further taps on potentiometers 69 and 73 via diodes 75 and 76, which lead back to the negative input of the operational amplifier 6o, the overall limit of the controller for the maximum pulse lengths.

The D component of controller k-5 is at the input for the Actual value formed (capacitor 55 and resistor 56). The output voltage of the regulator is fed to a summing amplifier 85 with the output of the threshold switch 38 and 89 returned signal compared and the differential voltage finally reaches the Switch the three-point switch with the two threshold values 38 and 89. The so-called dead 3one of this treipoint switch ers is determined by the resistance combination 91, 90 and 92 determined and the hysteresis by the Choice of resistors 100 and 96 as well as 10 'and 9 ^ · The two switching outputs of the threshold value switch 88 and 39 control the two output stages with field effect transistors 106 and 107 and the two! Actuator simulation with the correct sign. The decay time of the timers is determined by the diodes 111 and 113, resistors 110 and 1: 2 as well the capacitor 120. The decay time is oriented

.1 7: i J 7 9 Π

read on ie rl works?! jndensator 3 120 and resistance "1? With this resistance having a significantly higher value Has value than the two resistors 110 and 112. The simulation signal comes from the timer via the Impedance converter with the operational amplifier 11p and via a weighting resistor 116 to a summation point in front of your negative input of the operational amplifier 35, iem also the output signal of the PID controller is fed.

The following points of the subject of the application can be summarized point out..

The PID controller ^ 5 and the three-point switch U6 have a limitation of the manipulated variable; is at the same time
the switching time is defined by means of switching hysteresis and the time constant en of the returned actuator simulation. With these measures, it is possible to stabilize the control loop at all operating points, even with variable route saving providers.

The} rö.3e of the dead time is three point se hold he 3 is determined by the desired accuracy and by the gain.

The size of the switching hysteresis is based on existing interference band and is also used to define the time of suspension.

The feedback of the control pulses, i.e. the output signal the three-point switch s ^ 6, takes place with different Time constant for the switching and reverse torque and influences the switching frequency.

3933290

The minimum pulse length is influenced the pecking in the Sinschaltmonent soie luror the enforcement of repatriation. You Oemi2o after yourself the smallest, which just barely have a mechanical effect Pulse length and thus forms the smallest correction step.

The maximum pulse length is smaller than would be sufficient for continuous oscillation. she will determined by as a limitation of the controller output.

The side behavior of the PID controller U5 is independent on the size of the control deviation. The? -7 gain is calculated according to the requirements of stability, the I component of the accuracy within the dead zone, with between the required damping and a compromise has to be made with regard to the control speed.

It applies to the limitation that the I component is just outside the dead zone is limited, which is useful against so-called low-frequency sawing. We are excited! the P gain is also reduced. To limit With the maximum pulse length, the controller is completely limited further away from the dead zone (See Figures 3 and M. The limitation of the I component prevents the I controller from running due to dead time behavior of the controlled system or large delays runs far away, causing continuous oscillation to lead. would. Because of this, the I-3 area becomes straight chosen so large that the accuracy is guaranteed is.

Blank page

Claims (1)

31 .3. 1932 Mü /? I ROBERT BOSCH GMBH, 7OOO STUTTGART 1 Expectations M. ) Device for exhaust gas recirculation in an internal combustion engine that works in particular with compression ignition "Schine, with a valve in the return line, the position of which can be determined by activating two solenoid valve windings (U1 and U2), and with a control device which, based on operating parameters, forms the control signal for the two solenoid valve windings, characterized in that the control device has an? ID -Has control behavior with limitations and the exhaust gas recirculation valve (13) can be controlled via the two solenoid valve windings (h 1 and h2) in three-point operation. 2. Device according to claim 1, characterized in that there are at least two of the larger operating parameters as operating parameters. Accelerator pedal position, speed, air flow ζ ir. Intake pipe, injection duration and temperature can all be processed. 3. Device according to claim 1 or 2, characterized in that the three-point switch enabling the three-point operation of the feedback valve (13) it has a hysteresis having. 3 2 3 J ι β 0. -i. Device according to claim 3, characterized in that In general, there is a signal processing unit with the three-point controller '^ d) has a return. 5. Device according to at least one of claims 1 "to ~, characterized in that the switching pulses for iie excitation windings (^ 1 and -2) between one minimum and a maximum value, which are related to the response behavior of the exhaust gas recirculation valve (13) as well as the overall behavior of the system. 6. Einri: htur.g according to at least one of claims 1 to 5, characterized in that the proportion of the returned Exhaust gas can be regulated to a target value that is based on the required amount of fresh air and this amount of fresh air based on load and speed determined w i r d. ". Device according to claim 6, characterized ± &S> is used as a Einspri Lastsigrial a fuel zdauer signal ~ 3ventils use, said Spritziauersigr.al is preferably corrected speed dependent. 3. Device according to claim T, characterized in that the load signal is a correction signal from the accelerator pedal is superimposed and this corrected signal is the input variable of a lambda characteristic curve (32) Output an air volume signal can be removed. 9. Device according to at least one of claims J to 3, characterized in that a PID controller -5, separate limits for the I and P components having. 10. Device according to at least one ier claims 1 'tis 9 j characterized in that the return Has delay behavior. 11. Device according to claim 10, characterized in ", that the time constant of the feedback is longer than is the decay time constant.